I would like to weigh in on this one as well.
There is some research out there on undetected offences. First of all, sexual assault is the most under-reported crime. People report their vehicles stolen and their houses broken into, but for a wide range of reasons, a lot of victims won't come forward when they've been sexually assaulted. It's very under-reported. That's the first thing.
There has been some research on undetected offences by offenders in treatment. There's one research project where they found that 23 offenders had an average of about 175 victims each for whom they had not been apprehended. Another study indicated that 232 child molesters had admitted more than 55,000 incidents of molestation. I have another study here by Gene Abel, who is well known in this area, which found that 561 offenders—a broad range of sex offenders, not just one type—had admitted to 291,000 offences. It's well known, certainly in our law enforcement circles, that often, but not always, you're dealing with the tip of the iceberg, and that more comes out through treatment, because there's perhaps no longer any jeopardy at that stage of the game for the offender.
The other thing that has been commented on in Canada by well-known researchers is that, as you indicated, using official records as a measure of recidivism is weak. Usually the follow-up periods are four to five years, which is a relatively short period. The offender may have been incarcerated for part of that time. The other thing is that you're dealing with a criminal conviction. So, first, a lot of the time, you don't get caught. Then, second, you have to be charged. Third, you have to be convicted. There are several processes before you get to the end of the road. Measuring this solely by using conviction statistics has its limitations.
I asked our research and development section to measure the effectiveness of sex offender registries. In their search around the world, they could not find any research measuring the effectiveness of an SOR, but it's a project that the RCMP will be undertaking. We already have a project to measure recidivism of offenders in the national sex offender registry. About 18% of offenders currently in the national sex offender registry have a prior conviction for sexual assault.